Probabilistic environmental risk assessment of non-dissolvable nanomaterials (TiO2, CNTs, graphene-based materials,...

The accumulation of engineered nanomaterials and nanoplastics in soils, especially sludge-treated soils, raises concerns regarding their potential environmental hazard and risks. This work applies probabilistic environmental risk assessment based on probabilistic species sensitivity distributions (pSSD) and published data from exposure models to derive risk characterization ratios (RCR) for four non-dissolvable nanomaterials: TiO2, carbon nanotubes (CNTs), graphene-based materials (GBMs), and polystyrene nanoplastics (PS-nanoplastics). Median predicted no-effect concentration (PNEC) values were found to be 36, 1.5, 56, and 0.5 mg.kg-1, for TiO2, CNTs, GBMs and PS-nanoplastics, respectively. No statistically significant differences were found between the effects of nanomaterials on plants versus invertebrates, or on monocotyledonous versus dicotyledonous plants. Form-specific comparisons also showed no statistical difference among the different nanomaterial forms (TiO2: P25 vs. anatase vs. rutile; CNTs: multi-walled CNTs vs. single-walled CNTs; GBMs: graphene vs. graphene oxide vs. reduce graphene oxide). At currently predicted environmental concentrations in European soils, none of the materials exhibits an environmental risk, except for TiO2-NPs in sludge-treated soils, where 66% of the simulations resulted in a RCR higher than 1 (median RCR = 1.38). However, a projected increase in environmental concentrations over time can result in larger potential risks in the future. This work also highlights the limitations of the current state of the art in terms of the availability of environmentally relevant ecotoxicity data in soils and of realistic predicted environmental concentrations.